Summary: A Framework for Index Bulk Loading and
Dynamization
Pankaj K. Agarwal ? , Lars Arge ?? ,
Octavian Procopiuc ? ? ? , and Jeffrey Scott Vitter y
Center for Geometric Computing, Dept. of Computer Science,
Duke University, Durham, NC 27708­0129, USA.
fpankaj,large,tavi,jsvg@cs.duke.edu
Abstract. In this paper we investigate automated methods for exter­
nalizing internal memory data structures. We consider a class of balanced
trees that we call weight­balanced partitioning trees (or wp­trees) for in­
dexing a set of points in R d . Well­known examples of wp­trees include kd­
trees, BBD­trees, pseudo­quad­trees, and BAR­trees. Given an efficient
external wp­tree construction algorithm, we present a general framework
for automatically obtaining a dynamic external data structure. Using this
framework together with a new general construction (bulk loading) tech­
nique of independent interest, we obtain data structures with guaranteed
good update performance in terms of I/O transfers. Our approach gives
considerably improved construction and update I/O bounds for e.g. ex­
ternal kd­trees and BBD­trees.
1 Introduction

Source: Arge, Lars - Department of Computer Science, Aarhus Universitet

Collections: Computer Technologies and Information Sciences